Chapter 10

Silanes and Siloxanes as Coupling Agents

to Glass: A Perspective

Janis G. Matisons

10.1 Composites and Coupling Agents

Silicate glass-reinforced composites, based on synthetic resins such as phenolics,
ureas, epoxies, melamines, and unsaturated polyesters, generally became available
in the 1940s. The importance of such products in a number of areas, such as aircraft
and marine applications, was soon realized, as these products offered high strength
and modulus at a reduced weight. However, it was soon uncovered that such prod-
ucts were very susceptible to the effects of ambient humidity. Although the specific
dry strength and modulus of these reinforced composites exceeds that of aluminum
or steel, upon prolonged exposure to humidity, a dramatic decrease in these proper-
ties is seen in such environments [1, 2].
Furthermore, the coefficient of thermal expansion for the glass fiber is very much
lower than that of the polymer resin. Thus, if the resulting polymer composite is
exposed to extreme temperature cycling, the difference between these thermal ex-
pansion coefficients results in stresses at the interface between the organic polymer
and the inorganic glass. Such stresses at the interface may even exceed the strength
of the composite [1, 2].
Commercial glass fibers used in reinforced composites are almost always pre-
treated with a coupling agent, which is capable of interacting with both the organic
polymer resin and inorganic oxide substrate. Such a coupling agent must not only
ensure that the physical properties of the reinforced material remain relatively unaf-
fected by moisture or humidity, but must also reduce the stress at the interface during
excessive thermal cycling. Trialkoxysilanes, which contain organic groups compat-
ible with the polymer resin, are the most commonly used coupling agents. The ad-
dition of almost any trialkoxysilane coupling agent to the glass fiber surface, will
improve the water resistance of the resulting reinforced composite. However, it is
important to note that such silanes are usually applied from aqueous solution, where

J.G. Matisons ()

Gelest Inc., 11 East Steel Road, Morrisville, PA 19067, USA
e-mail: jmatisons@gelest.com

M.J. Owen, P.R. Dvornic (eds.), Silicone Surface Science, Advances in Silicon Science 4, 281
DOI 10.1007/978-94-007-3876-8_10, © Springer Science+Business Media Dordrecht 2012
282 J.G. Matisons

Fig. 10.1 Mechanism of silane coupling to surfaces. M is a mineral element (Si for glass)

both hydrolysis and condensation of the silane coupling agent occur (see Fig. 10.1),
resulting in the formation of oligomeric siloxane materials (oligomerization takes
place in solution several hours prior to the silane being applied to the glass). Such
oligomers may well be what is actually adsorbed onto the glass surface, given that
the silane may stand awaiting application for several days. Application of silane
coupling agents from an aqueous solution therefore represents a dynamic process,
which is highly dependent on the age of the solution being applied. Thus, there are
a number of factors affecting the reproducibility of the coupling agent application,
and so affecting the final properties of the composite (its physical properties and
water resistance). For a particular silane coupling agent, the main factors affecting
its final properties relate to its degree of oligomerization and cross-linking.
These properties are more easily controlled in the glass fiber industry if the silane
is supplied very shortly after manufacture, and thus is not partially oligomerized
and/or cross-linked prior to use. Coupling agents, as supplied to the glass fiber in-
dustry, can undergo varying degrees of oligomerization and cross-linking. There
is a need for an aqueous coupling agent solution which remains stable, or at least
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 283

constant with respect to its state of oligomerization, no matter how much time has
lapsed since its manufacture.

10.2 The Glass–Polymer Interface

The deleterious effects of water on the mechanical properties of many metal oxide or
glass-reinforced composites are well documented [1–5]. Diffusion of and interaction
with water at the filler-polymer interface is responsible for the delamination between
the glass fiber and the polymer matrix. To overcome such problems, coupling agents
are used to generate a water resistant interface between the polymer and the glass
or for that matter, any inorganic filler [1, 2, 6–8]. These coupling agents must be
able to react or interact with both the glass surface and the polymer, to improve the
overall performance of the final reinforced composite materials [1, 2, 8–12]. It has
been established that if only a small amount of silane coupling agent is added to an
inorganic filler, the performance of the resultant composite will improve [1, 2]. The
site-selective adsorption of silanes and their oligomers at predefined positions on
solid surfaces is a key fabrication step, and a major challenge in many applications.
There have been several theories proposed to explain how silane coupling agents
improve composite performance. The variety of applications for the trialkoxysilane
coupling agents precludes any single theory used to explain their effectiveness in
improving the composite properties. The chemical bonding theory, where a silane
coupling agent formed covalent bonds with both the polymer resin and the inor-
ganic substrate, was proposed independently by Arkles [13] and Plueddemann [14].
Investigating some 142 silanes in epoxy- and polyester glass laminates, Pluedde-
mann [14] found that the overall composite properties greatly improved when a
silane was used that could chemically react with both the resin and the substrate.
A conventional glass fiber sizing solution always contains more than just the
silane coupling agent. In fact the silane coupling agent is not even the major active
component of such a sizing solution. Many sizing solutions contain more than a
dozen different active chemicals. In such complex formulation chemistry, generally
resident to the patent literature, it is often difficult to unravel what is really important
in making sizing formulations effective.
Basically a sizing agent for glass fibers must contain at minimum the following
active ingredients:
• 4–7 % film forming agents (polymers)
• <1 % silane coupling agents
• ∼0.1 % lubricant or mixture of synergistic lubricants
• <0.1 % electrostatic agent
• remainder is water (generally as an emulsion, and considered an ‘inactive’ ingre-
dient)
As the molten glass is drawn from the furnace into fibers it cools rapidly. On cool-
ing, the glass sizing solution (containing the silane coupling agent) is sprayed onto
the cooling fibers in a confined space. The fibers in this space move at several tens of
284 J.G. Matisons

meters per second, and without the addition of a film former, the fibers are brittle and
can rupture. While the hydrolysis and condensation of the silane takes place the film
former maintains the integrity of the cooling fiber. Clearly, in these circumstances,
the picture of the hydrolysis and condensation of a sole coupling agent to the glass
fiber surface (Fig. 10.1) is a simplification of a more complex system. Nevertheless,
this is the mechanism for surface treating such fibers. It provides an initial approach
to understanding the mechanism for surface treating such fibers as well as simpler
(in terms of monitoring the surface chemistry), non-heterogeneous surfaces such as
silica.

10.2.1 Silane Hydrolysis and Condensation

Trialkoxysilanes, with the general formula RSi(OR )3 , where R is a functional group

similar to, or compatible with, the polymerizing functional group of the polymer
resin, and R is a hydrocarbon radical (usually methyl or ethyl), are generally used
in composites manufacture. Alkoxysilanes are applied from dilute aqueous solu-
tions, as partial hydrolysates, or from organic solvents (generally alcohols) [1, 2,
15]. All such silane coupling agent solutions undergo initial hydrolysis and some
oligomerization, prior to interacting with glass substrates. Initially, such silane cou-
pling agents may interact with glass surfaces through hydrogen bonding with the
glass-surface hydroxyl groups. Subsequently condensation of these initial surface
structures generates siloxane bonds to the surface (Fig. 10.1). It is also possible
that some lateral polymerization occurs without the formation of bonds to the sur-
face [16]. Irrespective of how such a siloxane film is formed on the substrate, it
generally consists of multiple siloxane layers [17–19].
Siloxanes are generated from chloro- or alkoxysilanes under hydrolytic condi-
tions, which involve silanols as reactive intermediates [9]. To selectively obtain a
specific siloxane-surface structure, it is crucial to control the competition between
silanol formation and silanol condensation [6]. The presence of a sterically demand-
ing group on the silane can successfully slow down the condensation reaction, and
thus permit the formation of stable silanols, silanediols, and silanetriols [20–24].
The primary condensation products of di- or tri-functional silanols such as disilox-
ane 1,3-diols or disiloxane 1,2,3-triols are in fact functionalized oligomeric silox-
anes themselves [25, 26].
Organosilanes containing various organic groups, such as alkyl [27–30], per-
fluoroalkyl [31], phenyl [32], and vinyl [33] groups, have been used for the sur-
face modification of layered silicates. Silylation is now also common for the im-
mobilization of organosilyl groups onto layered titanates [34, 35]. Such diversity,
however, is not apparent when it comes to glass surfaces, where traditionally used
silanes still occupy >99 % of the commercial applications. Of these silanes, γ -
aminopropyltrialkoxysilane is used in well over 60 % of the treated glass fiber mar-
ket; while γ -glycidoxypropyltrialkoxysilane and methacryloxypropyltrialkoxysi-
lane also maintain significant use.
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 285

Oxane bonds (see Fig. 10.1) that form between silane coupling agents and any in-
organic substrate are easily hydrolyzed [36, 37]. However, this hydrolysis and bond
re-formation remain in true equilibrium, and hydrolyzed oxane bonds will readily
re-form [36, 37]. Improved composite properties occur where hydrolysis and con-
densation reactions are in equilibrium. It has been suggested that these hydrolysis
and condensation reactions provide a mechanism for stress relief at the interface
[1, 2]. As a consequence, a silane/siloxane interphase forms at the surface.

10.2.2 Factors Affecting Silane Adsorption

There are a number of factors which influence the structure of the silane coupling
agent interphase. Firstly, the pH of the aqueous silane solution is important, since
basic or acidic conditions affect the relative rates of silane hydrolysis and condensa-
tion [9]. The condensation of neutral alkoxysilanes with glass and silica is catalyzed
by the addition of aliphatic amines [18]. The tensile strengths of the composites
made from these catalyzed silanes are greatly improved [1, 10, 18]. Acidic or basic
conditions are also found to increase the amount of silane adsorbed [10, 18]. The
surface potential of the oxide substrate also varies with the pH of the applied solu-
tion, affecting the orientation of the adsorbed silane layers [19]. This effect of pH
upon surface potential is more complex on mixed-oxide substrates, such as glass,
where surface micro-heterogeneities exist, such that the resultant surface potential
is not a simple average of the component oxide potentials [19].
The selection of the trialkoxysilane coupling agent may, in fact, contribute to the
poor water resistance properties of the composite, as the treated substrate is still
hydrophilic [37–40]. This is certainly the case for amino-functional silanes, where
excess amine still exists at the interface. Such hydrophilicity has been countered by
either (i) using very-dilute silane solutions, or (ii) by washing the treated surface
with solvent (water or the appropriate organic solvent) to remove any excess, non-
covalently bound (or physisorbed) silane [41–48]. Alternatively, a mixture of amino-
functional silane and phenyltrimethoxysilane can be used to impart a high degree of
hydrophobicity to the resultant surface [37, 38].
Basic functional groups such as amines will self-catalyze the hydrolysis reac-
tion leading to more aggressive monolayer formation as compared to non-animated
silanes [44]. The initial hydrolysis step can occur either in solution or at the substrate
surface depending on the amount of water present in the system. An overabundance
of water will result in excessive polymerization in the solvent phase, while a defi-
ciency of water will result in the formation of an incomplete monolayer.
The drying conditions used for the silane treated substrate also affect the structure
of the adsorbed silane [49, 50]. The temperature and duration of the drying proce-
dure will influence the number of siloxane bonds formed between adjacent silanes
(siloxane formation) as well as with the surface [49, 50]. The generation of a silox-
ane coating with multiple surface bonds results in improved composite performance
[41, 43, 44, 49, 50]. This is because the probability of all the siloxane-surface bonds
286 J.G. Matisons

being hydrolyzed at the same instant in time is remote. Solvent, concentration, reac-
tion time, and reaction temperature all have an effect on the attachment kinetics, but
most studies only examine one or two of these parameters, and so often a consistent
picture of silane adsorption remains missing.

10.2.3 Silane–Polymer Interactions

The oligomeric silanols formed from the hydrolyzed silanes, and attached to the sub-
strate, must retain some degree of solubility/compatibility in order to interact with
the polymer resin [41, 43]. If both the oligomeric siloxane layer and the polymer
resin are compatible, a copolymer can result upon cure. However, if the oligomeric
siloxane and polymer resin are only partially compatible, the resin and the siloxane
will cure separately, generating an interpenetrating polymer network of the cou-
pling agent residing on the substrate and within the polymer matrix [7, 49, 50].
Pseudo-interpenetrating polymer networks result from the weak secondary bonding
interactions between the oligomeric siloxanes and pre-formed thermoplastic poly-
mer resins, as here only the silane can cross-link through the formation of condensed
siloxane bonds [7, 50].
Silane coupling agents lower the surface tension of a substrate, wet it and make
its surface energy higher, and hence accessible for effective bonding [13]. Thus,
a hydrophobic matrix (resin composite) can adhere to hydrophilic surfaces, It is
now well established that more than a monolayer of silane coupling agent is re-
quired on the substrate in order to optimize the strength of the resultant composite
[1, 2, 7]. In fact, an optimum thickness of coupling agent must be achieved in or-
der to obtain optimal overall performance of the composite [1, 7, 49]. The amount
of γ -methacryloxypropyltrimethoxysilane adsorbed upon E-glass fibers, affects the
curing process of a vinylester resin at a far greater distance than the thickness of the
silane interphase [45, 50, 51]. Previous work has shown that excessive amounts of
a silane at the interface resulted in a reduction in the composite fracture toughness,
due to the final resin becoming brittle [51].
The highly flexible polymeric siloxane backbone, arising from the silane hydrol-
ysis, enables the interphase to adjust to steric constraints imposed by the oxide filler
surface. Furthermore, the ratio of hydrophobic to hydrophilic groups may be ad-
justed, by either using more than one silane, or by using a hydrophobic hydrocarbon
chain to adjust the distance that a polar hydrophilic group maintains from the sil-
icon atom. In this way the number of polar group interactions with the polymer
resin can be optimized, such that the polymer “sees” a continuous reactive surface
on the fiber or filler, which results in maximum dry strength and durability in the
resultant composite [7, 50, 52]. It is, therefore, necessary to control the hydrolysis
and oligomerization rates very carefully, if controlled and reproducible silane mod-
ified surfaces are to be produced. It is also necessary to control the degree of silane
cross-linking (through condensation) and size of the polymeric siloxane segments,
to ensure that their interpenetration into the polymer matrix results in optimum com-
posite properties [8, 45, 50].
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 287

Usually, the surface treatment is carried out with a silane water–alcohol solution
in concentrations of 0.5–2 % by weight. These conditions offer several advantages,
in particular
(i) an increase of silane solubilization,
(ii) better control of the surface film thickness, and
(iii) more uniform surface coverage.

10.2.4 Acid-Base Perspectives

Fowkes and coworkers first described the interaction between polymers, fillers and
silane coupling agents in terms of their respective acid-base properties [53–55]. Us-
ing the principles first described by Drago [56], they characterized these materials as
either Lewis acids or bases, from calorimetric and/or spectroscopic measurements.
Such information was then used to explain the interactions between the materials
produced, which affected their solubility, wettability, adsorption and adhesion prop-
erties [53–55]. For example, the acid-base nature of various silane treated fillers
affects their dispersion in a range of polymers, as well as the viscosity of the final
mixtures. The orientation of some silane coupling agents on the surface has simi-
larly been explained in terms of the respective acid-base properties of the silane and
the substrate. Employing angle-resolved X-ray photoelectron spectroscopy (XPS)
and zeta-potential measurements, Fowkes found the methacryl functional group
in γ -methacryloxypropyltrimethoxysilane, and the amino-functional group in γ -
aminopropyltrimethoxysilane were both oriented towards the surface of a magne-
sium aluminum silicate glass powder [53–55].
The role of acids or bases in the adsorption of silanes or siloxane polymers, espe-
cially if applied from organic solvents, cannot be overlooked. Leyden and coworkers
[57, 58] investigated the interactions between trimethoxysilane, HSi(OMe)3 , and
Cab-O-Sil in toluene, in the presence of various amines. They concluded that all
amines catalyze the interaction between the silane and the surface silanols of silica;
however, amines with exchangeable protons do have an additional catalytic effect.
The presence of boron on silica surfaces is known to enhance the reactivity of
surface silanol groups [60]. Elevated levels of boric oxide in E-glass formulations
were found to enhance silane adsorption on such surfaces [61]. Similarly, silica
surfaces treated with boron trichloride, followed by washing with water, produce
B-OH surface groups on silica [61]. These B-OH groups are more reactive, than
Si-OH groups towards trialkoxysilanes [60].
The structure and dynamics of alkoxysilane chemisorption onto metal oxides
and glass was studied by many techniques including nuclear magnetic resonance
(NMR), Fourier-transform infrared spectroscopy (FTIR), XPS, streaming zeta-
potential and secondary ion mass spectrometry (SIMS) [42, 50, 53–55, 62–66]. The
nature of the substrate selected helps determine whether the nature of the chemisorp-
tion process is easily identified. Also, the silane-substrate system under study deter-
mines which spectroscopic technique will reveal the most about the chemisorption
288 J.G. Matisons

process. It is often useful to attach “identifying groups” to the silane, or ensure that
the silane selected is likely to undergo chemical reactions with the surface which
may be followed spectroscopically, in order to achieve a better understanding of
the chemisorption processes. Unfortunately, the chemistries of the most industrially
useful silanes and substrates are not always amenable to such spectroscopic tech-
niques.

10.3 Surface Structure and Adsorption Processes

There has been little theoretical treatment of real surfaces, which are both non-
uniform and non-planar. The impact of such surfaces on polymer physisorption has
usually been left up to experimentalists. There have also been relatively few system-
atic studies of the effects of chemical heterogeneity [65, 66]. Physical heterogeneity,
and in particular the geometry of the surface, has received more attention.
Although most theoretical treatments assume planar geometry, there have been
some investigations on the adsorption of polymers on spherical particles [9, 40, 50,
60]. It is predicted that the effect of the curvature of the surface is more pronounced
as the radius of the particles approaches that of the polymers (i.e. the radius of
gyration, rg ). The thickness of a layer of poly(vinyl alcohol) (Mw = 67,000, rg =
11.7 nm), adsorbed from water onto polystyrene latex particles decreased by a factor
of two, when the size of the latex particles decreased from 250 to 50 nm [40]. The
effects of pore size on polymer adsorption were examined and it was concluded that
adsorption also increased with pore size [40].

10.3.1 Adsorption on Silica Surfaces

Pure silica surfaces dominate the studies of adsorption. There are a number of rea-
sons why the majority of work on adsorption has been conducted using silica as
the substrate. This homogeneous substrate’s surface properties have been well char-
acterized, to the extent that it is possible to quantitatively follow adsorption by a
number of spectroscopic as well as non-spectroscopic techniques [37, 38]. For ex-
ample, there are two main types of surface silanol groups on silica, the isolated and
the vicinal (which are those within close proximity to one another), which are easily
distinguished by FTIR spectroscopy [38]. The surface area of the various types of
silica, and the numbers of isolated and vicinal silanol groups per 10 nm2 , have also
been well established by numerous techniques [38, 39]. It has, therefore, been pos-
sible to follow silane adsorption and to detect if it will occur at the isolated silanol
functional group, as the FTIR band of this group decreases proportionally with the
degree of silane adsorption [38]. FTIR studies of pyridine and ammonia adsorption
also confirm that the isolated silanol is the adsorption site for such small molecules
[38]. Furthermore, infrared analysis of the exchange between D2 O and the surface
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 289

silanol groups on silica identifies the vicinal silanol groups as sites for water adsorp-
tion.
XPS is able to very precisely determine the elemental composition on the sub-
strate surface, and so remains an important technique for determination of the extent
of surface coverage on glass, as the minor elements in the glass can reveal incom-
plete silane surface coverage. Sum frequency generation (SFG) (see Chap. 2) has
proved a new and exciting addition to monitoring surface silane treatments. SFG
spectroscopy is a second-order nonlinear vibrational technique, with an intrinsic sur-
face selectivity and sub-monolayer sensitivity [58], known to be more effective than
conventional vibrational techniques, such as infrared and Raman [38, 59], for re-
vealing the structural ordering, arrangement, and composition in the organic mono-
layers adsorbed on various substrates, including oxide surfaces [58]. Furthermore,
SFG offers the only means to study the interfacial water structure.
Shafrin and Zisman investigated the effects of relative humidity on silicate glass
surfaces by contact angle measurements using methylene iodide (a non-hydrogen
bonding organic liquid) [93]. They measured a contact angle of 13° at 1 % relative
humidity (RH), and a contact angle of 36° at 95 % RH [93], similar to the contact
angle of 37° for methylene iodide on water. These results may be better understood
when compared with that observed for methylene iodide on a clean silica surface,
where in an ultra-high vacuum (i.e., where silica has only surface silanol groups), a
contact angle of ∼10° is measured [93]. If water vapor is slowly admitted into the
vacuum chamber containing the silica sample, the fractional monolayer coverage
of molecular water gradually increases, and contact angles from 11°–20° are suc-
cessively measured. Above 20°, however, adsorbed water forms multilayers on the
silica sample. Therefore, Shafrin and Zisman concluded that at high RH, multilayers
of water are adsorbed on the silicate glass surface; whereas at low RH, only resid-
ual amounts of adsorbed water are present. Subsequent adsorption isotherm studies
have confirmed this conclusion.
A composite isotherm was produced from a number of studies in different pres-
sure regions [93–95]. Silicate glasses subjected to RH levels between 1–50 % grad-
ually form a monolayer of adsorbed molecular water. Above 50 % RH, multilayers
of water adsorb until a thick film forms. Placing the monolayer films under low
pressure (between 10−3 to 100 Torr, which corresponds to 0.005–5 % RH), the re-
sulting isotherms indicate that only a small fraction of an original water monolayer
remains. It is, however, unclear whether this residual moisture is due to molecular
water, or surface silanols, which may be thought of as chemisorbed water.

10.3.2 Adsorption on Heterogeneous Surfaces

The adsorption processes on other homogeneous metal and metal oxide surfaces
have also been studied, and often the nature of the interaction was identified as be-
ing similar to that of silica [37, 38, 69–71]. The only extensively examined hetero-
geneous surface is glass. In particular, the chemisorption of silane coupling agents
290 J.G. Matisons

onto E-glass fibers has received great attention because of its industrial relevance
[1, 2, 19, 95, 96]. In what follows, the main problems associated with all studies of
chemisorption upon heterogeneous surfaces are pointed out.
First of all, the existing chemical heterogeneity may not only vary from manu-
facturer to manufacturer, but also with the history of the substrate, i.e. how it was
cleaned and stored. Thus, in making comparisons between studies of chemisorption
on similar substrates, such as E-glass fibers, it is important to know the exact sur-
face chemistry of the substrates being compared. There is also the possibility that
chemical micro-heterogeneities may exist, which complicate the matter further [19].
Adding sodium, boron, calcium, and alumina to glass shifts the surface isoelectric
point of a quartz glass or pure silica to a higher pH. The consequences in terms of
absorption are dramatic. For instance, carboxylic acids will not absorb on silica, but
will on E-glass fibers that have trace amounts of boron and alumina.
Secondly, there are a number of geometric forms in which the same sample may
be presented, for instance plates (microscope slides), cylinders (fibers) and spheres
(powder), and chemisorption of silanes on a substrate may be substantially differ-
ent between them. Furthermore, chemisorption may also be affected by differences
in diameters for the same geometry [40]. The surface area of the substrate is of
vital importance. This will influence not only the chemical interaction of the sur-
face (as more surface functional groups promote better interaction with adsorbed
molecules), but also the sensitivity required by the analytical technique employed to
monitor the chemisorption process (the smaller the surface area the greater the sen-
sitivity required). For these reasons, studies of chemisorption of small molecules,
such as silane coupling agents, have generally focused on large surface area, homo-
geneous particles, such as silica.
When considering the adsorption of polymers onto surfaces, there is always the
possibility of patch-wise adsorption. This type of adsorption process has been exam-
ined using a multifractal approach [72]. Chemical and/or physical heterogeneities
on the surface may be responsible for patch-wise adsorption, which is thought to
proceed in a multi-step growth process. Each step occurs with smaller and smaller
probability, but results in bigger and bigger patches. The interplay between these
two events results in fractal behavior, a fractal being defined as a geometrical struc-
ture with an irregular or fragmented appearance. A multifractal approach is required
when the interplay between the chemical and physical heterogeneities of the surface
is considered.

10.4 Glass Surfaces

The composition of a glass will vary with its intended application [73–76]. Sodal-
ime glass (see Table 10.1), composed primarily of oxides of silicon, sodium and
calcium, is commonly used for bottles and containers. Pyrex, a borosilicate glass,
has high resistance to thermal shock due to the presence of boron oxide, making it
suitable for laboratory and kitchen glassware. E-glass fibers, the most common type
of glass employed in textiles or reinforced composites, are also borosilicates. S-glass
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 291

Table 10.1 Constituents of

commercial glasses by Component Sodalime Pyrex (%) E-glass (%) S-glass (%)
weight % glass (%)

SiO2 70–75 80–86 52–56 64–66

CaO 7–10 – 16–25 0–0.3
Al2 O3 0–1.5 0–2 12–16 24–26
B2 O 3 – 6–18 5–10 –
MgO 0–4 – 0–5 9–11
Na2 O 10–13 2–8 0–2 0–0.3
K2 O 0–1 – 0–2 0–0.3
TiO2 – – 0–0.8 –
Li2 O – 0–1 – –
SO3 0–0.5 – – –
Fe2 O3 0–0.2 – 0.05–0.4 0–0.3
F – – 0–1.0 –

fibers, however, are alumino-silicates, and are used primarily for high performance
materials which require fibers with very high tensile strength. Some minor oxide
ingredients are added not only for economic and production purposes, but also to
control and modify certain glass properties. Calcium and aluminum oxides control
or improve the expansion, durability and chemical resistance of the glass [73–76].
Alkali metal/alkaline earth oxides are added to reduce the melting temperature and
viscosity of the glass, by disrupting the continuity of the silica network (i.e. breaking
some of the Si-O bonds). However, alkali oxides also lower the chemical resistance
of the glass. The silica network is retained upon formation of the multi-component
silicate glass, and the non-bridging oxygen atoms are there to provide the necessary
charge balance for the added cations.
The surface concentrations of the various oxides, which comprise the glass, will
vary from the bulk composition, depending on the thermal history of the glass, the
relative humidity, and the surface treatment to which it was subjected after melting
and cooling [75–80]. The strength of a glass fiber is influenced by the nature of its
surface. Components which lower the surface free energy will diffuse towards the
glass surface (surface segregation), while the glass being formed is in its molten
state. Hydrolysis and leaching of the alkali and alkali earth metal silicates, and
volatilization of the alkali oxides (such as Na2 O and B2 O3 ) during glass melting
and cooling to room temperature, also affect the surface composition [77–83].
Immediately after glass manufacture, optical measurements have detected a
lower refractive index from a thin surface film. This very fine silica film is between
1–35 nm thick, and is due to the loss of alkali oxides by both volatilization, and the
hydrolysis/leaching of the alkali and alkali earth metal silicates [73–75]. Such thin
surface films have different chemical and physical properties from those of the bulk
glass, and help retard further bulk glass hydrolysis or leaching, by acting as a bar-
rier for component ion diffusion. Both thickness and density of the surface film vary
with glass composition, time, temperature and pH. A less durable glass produces a
292 J.G. Matisons

Fig. 10.2 (a) Conventional silane coupling agent and (b) Dipodal silane coupling agent (Matisons
and Kempson—unpublished data (1997))

thicker film than a more durable glass [73–75]. Studies in this author’s laboratories
[66] of the streaming potential of water sized E-glass revealed an acidic isoelec-
tric point (pH = 3) which, while consistent with that established for silica surfaces
(pH = 2.3), also indicated a shift to higher pH arising from the presence of added
alkaline components in the glass.
The drying conditions used for the silane treated glass also affect the structure of
the adsorbed silane/siloxane interphase. The temperature and duration of the drying
procedure influence the number of siloxane bonds formed between adjacent silanes
(siloxane formation) as well as with the surface. Such ‘siloxane coatings’ with mul-
tiple surface bonds (see Fig. 10.2) generally give improved composite performance.
The use of dipodal silanes has proved important in this respect.
Silanes are available in many forms, but two distinct structural types, the mono
and the bis, or dipodal silanes are most commonly used. Mono-silanes are of the
type R -(CH2 )n -Si(OR)3 , and bis- or dipodal silanes are of the type, (RO)3 Si-
(CH2 )n -R-(CH2 )m -Si(OR)3 (see Fig. 10.2). Mono-silanes are commonly used for
the organic functionalization of inorganic surfaces via condensation of hydrolyzed
silanols onto hydroxylated surfaces. These self-assembled monolayers (SAMs) do
not form highly cross-linked structures as silanols are mostly attached to the surface
and unavailable for the formation of siloxane bonds above the interface. Dipodal
silanes, however, with twice as many hydrolyzable alkoxy groups as mono-silanes,
form denser interfacial layers and stronger structures above the interface through
the formation of highly cross-linked siloxane networks.
Dipodal silanes are a new type of adhesion promoter because of their commer-
cial success in applications such as plastic optics, circuit boards, and on metallic
surfaces. These coupling agents are hydrolytically far more stable than conventional
silane coupling agents, yet have a significant impact on substrate bonding and com-
posite mechanical strength. Dipodal silanes show improved wet adhesion, improved
chemical resistance, good corrosion protection and improved composite processing.
Importantly, dipodal silanes also enhance film formation at the interface, where such
film formation is desired, as in corrosion protection. Film thickness is then related to
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 293

silane concentration and time of exposure to the silane solution. Dipodal silanes can
be readily mixed with conventional silane coupling agents to suit particular applica-
tions. The resulting siloxane film thickness that forms by introducing such dipodal
silanes depends mainly on the silane concentration and the substrates residence time
in the silane solution.

10.5 Sizing Formulations

It is well established that more than a monolayer of silane coupling agent is required
on the oxide substrate in order to optimize the strength of the resultant composite.
Only in this way is it possible to generate an interpenetrating network of the cou-
pling agent (resident on the fiber) within the polymer matrix. There is, however, an
optimum thickness of coupling agent which, if not achieved, results in a substan-
tial decline in the overall performance of the composite. A large flexible polymeric
backbone will enable the interphase to adjust to the steric constraints imposed by
the oxide surface and display a continuous reactive surface to the polymer matrix.
Furthermore, by using a mixture of silanes, the ratio of hydrophobic to hydrophilic
groups can be adjusted so as to optimize the number of polar group interactions
(which may also act as sites for water ingress) for maximum dry strength and dura-
bility [1]. Arkles [13] reported that it is necessary to control the hydrolysis and
oligomerization rates very carefully, if controlled and reproducible silane modified
surfaces are to be produced. It is also necessary to control the degree of cross-linking
and size of the polymeric silane segments, to ensure their interpenetration into the
polymer matrix for optimum composite properties.
Preformed siloxanes, like the silane products of hydrolysis and condensation, are
also capable of adhering to a variety of surfaces [39–45, 86–88]. They are strongly
water resistant polymers and should, in principle, also be able to give water resistant
interfaces between glass fibers and organic resins in composite materials. They dis-
play considerable backbone flexibility, so that they may also adjust to the availability
of the reactive sites on glass surfaces. Siloxanes may be synthesized with a variety
of functional groups attached, and the molecular weight distribution may be read-
ily controlled. The investigation of siloxanes bearing appropriate functional groups
may then lead to a whole new class of coupling agents, with all the advantages
of silanes, but with greater control and reproducibility of the surface modification
procedure. Importantly, siloxanes offer the prospect of combining the properties of
a polymeric film former (see Sect. 10.2) with that of a coupling agent in the one
molecule.
A number of factors affect the adsorption of polymers onto surfaces: these in-
clude the type of solvent; the polymer’s molecular weight and polydispersity index;
the concentration; the time allowed for an interaction; the number of reactive groups
per molecule, the functionality of reactive groups on polymers; the reaction temper-
ature; the pH in aqueous systems; the type of post-treatment; and the nature of the
substrate.
294 J.G. Matisons

Fig. 10.3 Siloxane-based

coupling agent (top)
Conventional
siloxane—PDMS (bottom)
(Matisons and
Kehoe—unpublished data
(1994))

Our group’s past research focused on siloxanes, which, like silanes, are strongly
water resistant polymers capable of binding tenaciously to a variety of surfaces, in-
cluding glass. Preformed siloxanes exhibit remarkable backbone flexibility enabling
them to be used as film formers, surfactants as well as coupling agents linking poly-
mers (having the appropriate functional groups) to glass surfaces. Their inherent
backbone flexibility, results from both the electronic and structural properties of the
Si-O and Si-C bonds, which permit unhindered rotation about the siloxane back-
bone in the case of PDMS. The freedom of rotation gives ideal screening for the
polar Si-O-Si backbone, by the non polar methyl groups, thereby giving the poly-
mer excellent film forming properties. As a result, siloxanes have very low surface
tensions usually between 20 and 25 mN m−1 , which promote their use as surfactants
in personal care products and in the textile industry.
Earlier studies in this author’s laboratories examined the attachment of a num-
ber of functionalized siloxanes onto E-glass fibers, and compared them to com-
mercial silane coupling agents [41–45, 86–88]. XPS and diffuse reflectance Fourier
transform infrared spectroscopy (DRIFT) were employed to establish the presence
of the siloxane on the glass surface; and to semi-quantitatively compare the modi-
fied surfaces. It was found that not only did siloxanes bearing trialkoxy-functional
groups (see Fig. 10.3) adsorb to glass fiber surfaces as effectively as a common
coupling agent, vinyltris(methoxyethoxy)silane; but, surprisingly, that other silox-
anes bearing a variety of functional groups (e.g. amino, aminohydroxy, hydrido, and
methacryl) also strongly adsorbed [41–45, 86–88].
The initial results with functionalized siloxanes prompted the synthesis and ex-
amination of a siloxane “coupling agent analog” bearing a large number of alkoxy
groups to E-glass fibers. This allows a comparison to be made among this siloxane
coupling agent analog (Fig. 10.3 top), the vinyltris(2-methoxyethoxy) silane used in
the earlier study, and a hydroxy-terminated poly(dimethylsiloxane) (Fig. 10.3 bot-
tom). DRIFT and XPS were again used to analyze the treated E-glass fibers [41,
43, 87]. However, the scanning electron microscopy (SEM) pictures in Figs. 10.4a
and 10.4b contrasting conventional sizing formulations (in aqueous solution) with
that of the pure siloxane coupling agent analog, applied in the case of the latter from
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 295

Fig. 10.4 (a): SEM of E-glass fibers sized with formulation containing film former, silane cou-
pling agent, lubricant, and antistatic agent (Matisons, J. and Kempson, S.—unpublished data
(1997)). (b): SEM of hydroxy-terminated PDMS sizing solution (Matisons and Le Huy—unpub-
lished data (1996))

Fig. 10.5 (a): Siloxane coupling agent (Fig. 10.3 top; m = 195; n = 3) applied from toluene
Solution (Matisons and Le Huy—unpublished data (1996)). (b): Siloxane coupling agent (Fig. 10.3
top; m = 175; n = 23) applied from toluene solution (Matisons and Le Huy—unpublished data
(1996))

toluene solution, illustrate more than anything else that such siloxanes can be used
to treat glass fibers.
An interpenetrating polymer network, IPN, that is formed between the glass
fibers and the siloxane coupling agent analogs when the pendant alkoxy groups are
increased and the siloxane film that forms is shown in Fig. 10.5a. Activated alkoxy
groups on the siloxane in this case turn to silanols that deposit onto the glass fibers
and during cross-linking not only form a siloxane film, but form polymeric bridges
between glass fibers as shown in Fig. 10.5b. Such cross-linked siloxanes have a
significant impact on bonding and mechanical strength.
296 J.G. Matisons

While the static mechanical performance benefits of fiber-reinforced compos-

ites are often the reason for their selection in structural applications, it is generally
accepted that the response of the fiber–matrix interphase region can contribute to
impact resistance and damage tolerance. The effect of the interphase on impact per-
formance is largely determined by the choice of sizing components applied during
glass fiber production. Published results indicate that the impact response of a fiber-
reinforced composite can be tailored towards high energy absorption by engineer-
ing weak fiber–matrix interfacial interactions; or, conversely, high damage tolerance
(e.g., residual strength after impact) can be produced by promoting strong fiber–
matrix interfacial interactions. These siloxane coupling agents uniquely lend them-
selves to not only protecting glass fiber integrity (i.e. damage tolerance), but also to
being able to absorb high energy impacts. As the fiber–matrix bond strength is in-
creased, energy absorption during impact decreases. Furthermore, the fiber-siloxane
coupling agent bond strength can now be tailored by adjusting the number of alkoxy
side groups on the siloxane backbone. A final note of caution should be added; mak-
ing siloxanes with alkoxy groups resident on each silicon atom along the siloxane
backbone is possible, but the utility of such siloxanes is very limited, as they gel
rapidly on contact with trace amounts of moisture, such as is resident on common
laboratory glassware.
In summary, pre-formed functionalized siloxanes containing alkoxy side groups
can be made to serve as effective sizing agents that combine the film forming prop-
erties of polymers, together with the coupling properties of conventional silane cou-
pling agents. It remains to optimize the molecular weight and the number of alkoxy
side groups on such siloxane coupling agents to generate the best possible sizing
results in commercial applications.

References
1. Plueddemann E (1990) Silane coupling agents, 2nd ed. Plenum Press, New York
2. Plueddemann E (1991) J Adhes Sci Technol 5:261
3. Bader MG, Bailey JE, Bell I (1972) In: Kriegel WW, Palmour H (eds) Ceramics in severe
environments. Material science research, vol 5. Plenum Press, New York
4. Atkins AG (1975) J Mater Sci 10:819
5. Outwater JO (1975) J Adhes 2:242
6. Brook MA (2000) Silicon in organic, organometallic, and polymer chemistry. Wiley, New
York
7. Angst DL, Simmons GW (1991) Langmuir 7:2236
8. Pantano CG, Wittberg TN (1990) Surf Interface Anal 15:498
9. Schubert U, Husing N (2000) Synthesis of inorganic materials. Wiley-VCH, Weinheim
10. Wang D, Jones FR, Denison P (1992) J Adhes Sci Technol 6:79
11. Allen KW (1992) J Adhes Sci Technol 6:23
12. Drown EK, Moussawi H, Drzal LT (1991) J Adhes Sci Technol 5:865
13. Arkles B (1977) Chemtech 7:766
14. Pleuddemann E, Clark H, Nelson L, Hoffmann K (1962) Mod Plast 39:136
15. Blum FD, Meesiri W, Kang H-J, Gambogi J (1991) J Adhes Sci Technol 5:479
16. Tripp CP, Hair ML (1991) Langmuir 8:1120
17. Arkles B, Steinmetz JR, Zazyczny J, Mehta P (1992) J Adhes Sci Technol 6:193
10 Silanes and Siloxanes as Coupling Agents to Glass: A Perspective 297

18. Hair ML (1986) In: Leyden DE (ed) Silane surfaces and interfaces, vol 1. Gordon & Breach,
New York
19. Ishida H (1983) In: Mittal KL (ed) Adhesion aspects of polymeric coatings. Plenum Press,
New York, pp 45–106
20. Jutzi P, Strassburger G, Schneider M, Stammler H-G, Neumann B (1996) Organometallics
15:2842
21. Simons R, Galat KJ, Rapp BJ, Tessier CA, Youngs WJ (2000) Organometallics 19:5799
22. Winkhofer N, Roesky HW, Noltemeyer M, Robinson WT (1992) Angew Chem 104:670
23. Ishida H, Koenig JL, Gardner KC (1982) J Chem Phys 77:5748
24. Al-Juaid SS, Buttrus NH, Damja RI, Derouiche Y, Eaborn C, Hitchcock PB, Lickiss PD
(1989) J Organomet Chem 371:287
25. Unno M, Alias SB, Saito H, Matsumoto H (1996) Organometallics 15:2413
26. Walawalkar MG (2003) Organometallics 22:879
27. Ogawa M, Okutomo S, Kuroda K (1998) J Am Chem Soc 120:7361
28. Ruiz-Hitzky E, Rojo JM (1980) Nature 287:28
29. Yanagisawa T, Kuroda K, Kato C (1998) React Solids 5:167
30. Okutomo S, Kuroda K, Ogawa M (1999) Appl Clay Sci 15:253
31. Ogawa M, Miyoshi M, Kuroda K (1998) Chem Mater 10:3787
32. Yanagisawa T, Kuroda K, Kato C (1988) Bull Chem Soc Jpn 61:3743
33. Isoda K, Kuroda K, Ogawa M (2000) Chem Mater 12:1702
34. Ide Y, Ogawa M (2003) Chem Commun 1262
35. Ide Y, Ogawa M (2005) Chem Lett 360
36. Abbenhuis HCL (2000) Chem Eur J 6:25
37. Lickiss PD (2001) In: Rappoport Z, Apeloig Y (eds) Chemistry of organic silicon compounds,
vol 3. Wiley, Chichester, pp 695–744
38. Iler R (1979) The surface chemistry of silica. Wiley Science, New York
39. Hariharan A, Kumar S, Russel T (1990) Macromolecules 23:3584
40. Cohen SM, Cosgrove T, Vincent B (1986) Adv Colloid Interface Sci 24:143
41. Britcher LG, Kehoe DC, Matisons JG, Netting AKO, Smart RStC, Swincer AG (1993) Lang-
muir 9:1609
42. Britcher LG, Kehoe DC, Matisons JG, Swincer AG (1995) Macromolecules 28:3110
43. Britcher LG, Kehoe DC, Matisons JG (2001) Polym Preprints 42:178
44. Arora PS, Matisons JG, Provatas A, Smart RStC (1995) Langmuir 11:2009
45. Britcher LG, Kehoe DC, Matisons JG, Arora PS, Smart RStC (1996) In: High performance
coating materials. Silicones in coatings, vol 1. Paint Research Association, New York, p 16
46. Jokinen AE, Matisons JG, Rosenholm JB (1997) J Colloid Interface Sci 194:263
47. Watson H, Jokinen AE, Mikkola P, Matisons JG, Rosenholm JB (1997) Prog Colloid Polym
Sci 105:80
48. Jokinen AE, Matisons JG, Rosenholm JB (1998) J Mater Sci Lett 17:149
49. Matisons JG (1998) In: High performance coating materials. Silicones in coatings, vol 2. Paint
Research Association, New York, p 5
50. Le-Huy CC, Britcher LG, Matisons JG (2002) Silicon Chem 1:195
51. Ikuta N, Suzuki Y, Maekawa Z, Hamada H (1961) Polymer 34:2445
52. Bell JP, Schmidt RG, Malofsky A, Mancini D (1992) J Adhes Sci Technol 5:927
53. Fowkes F, Dwight D, Manson J, Lloyd T (1988) Mater Res Soc Symp Proc 119:223
54. Dwight D, Fowkes F, Cole D, Kulp M, Sabat P, Salvati L Jr, Huang T (1990) J Adhes Sci
Technol 4:619
55. Dwight D, Fowkes F, Cole D, Kulp M, Sabat P, Salvati LJr, Huang T (1990) J Adhes Sci
Technol 4:690
56. Drago R (1973) Struct Bond 15:73
57. Blitz J, Murthy R, Leyden D (1987) J Am Chem Soc 109:7141
58. Shen YR (1984) The principles of nonlinear optics. Wiley, New York
59. Eggers PK, Da Silva P, Darwish NA, Zhang Y, Tong Y, Ye S, Paddon-Row MN, Gooding JJ
(2010) Langmuir 26:15665
298 J.G. Matisons

60. Pantano C, Carman L, Warner S (1992) J Adhes Sci Technol 6:49

61. Nitzsche S, Burkhardt J, Wegehaupt K (1971) Wacker-Chemie GmbH, German Patent, DE
1,955,514
62. Watson H, Mikkola PJ, Matisons JG, Rosenholm JB (2000) Colloids Surf 161:183
63. Watson H, Mikkola PJ, Matisons JG, Rosenholm JB (2001) Colloid Polym Sci 279:1020
64. Watson H, Norstrom AE, Matisons JG, Root A, Rosenholm JB (2001) J Adhes Sci Technol
15:1103
65. Britcher LG, Kehoe DC, Matisons JG (2000) In: Mittal K (ed) Silanes and other coupling
agents, vol 2. VSP, Utrecht, pp 99–114
66. Kempson SD, Matisons JG (1997) Chem Australia 12:3
67. Embery CJ, Clarke SR, Matisons JG (2003) In: Clarson S, Fitzgerald J, Owen M, Van Dyke
M (eds) Synthesis and properties of silicones and silicone-modified materials. ACS symp.,
vol 838. Amer Chem Soc, Washington, p 26
68. Matisons JG, Graser S, Britcher LG (2001) In: Mittal K (ed) Acid–base interactions—
relevance to adhesion, vol 2. VSP, Utrecht, p 601
69. Danner J, Vohs J (1992) Appl Surface Sci 62:255
70. van Ooij W, Sabata A (1993) Surf Interface Anal 20:475
71. Porro T, Pattacini S (1992) J Adhes Sci Technol 6:73
72. Vlad M (1993) J Colloid Interface Sci 159:21
73. Doremus R (1973) Glass science. Wiley, New York
74. Holland L (1964) The properties of glass surfaces. Chapman & Hall, London
75. Rosington D (1972) In: Pye L, Stevens H, La Course W (eds) Introduction to glass science.
Plenum Press, New York, p 101
76. Kruger A (1988) In: Nowotuy J, Du Four L-C (eds) Surface and near surface chemistry of
oxide materials. Materials science monographs, vol 47, Chap. 9, Amsterdam
77. Prabakar S, Mueller KT (2004) J Non-Cryst Solids 349:80
78. Fry R, Tsomaia N, Pantano CG, Mueller KT (2003) J Am Chem Soc 125:2378
79. Van Ginhoven RM, Jónsson H, Corrales LR (2005) J Phys Chem B 109:10936
80. Hall MM, Clare AG (2007) J Sol-Gel Sci Technol 41:107
81. Pedone A, Malavasi G, Menziani MC, Cormack AN, Segre U (2006) J Phys Chem B
110:11780
82. Kister S (1962) J Am Ceram Soc 45:59
83. Varshneya AK (1994) Fundamentals of inorganic glasses. Academic Press, London
84. Haaland D (1986) Appl Spectrosc 40:1152
85. Davydov V, Kislev A, Zhuravlev L (1964) Trans Faraday Soc 60:2254
86. Matisons JG, Provatas A (1998) Langmuir 14:1656
87. Ma R, Le-Huy CC, Britcher LG, Matisons JG (2001) Polym Preprints 42:248
88. Matisons JG, Provatas A (1994) Macromolecules 27:3397
89. Watson H, Norstrom AE, Mikkola PJ, Matisons JG, Rosenholm JB (2000) J Colloid Interface
Sci 232:149
90. Ooi K, Miyatake M (1992) J Colloid Interface Sci 148:303
91. Huang K, Balazs AC (1993) Macromolecules 26:4736
92. Kuksenok O, Yeomans JM, Balazs AC (2001) Langmuir 17:7786
93. Shafrin E, Zisman W (1967) J Am Ceram Soc 50:478
94. Paul AJ (1977) Mater Sci 12:2246
95. Clare AG (2007) In: Groza JR, Shackelford JF, Lavernia EJ, Powers MT (eds) Materials pro-
cessing handbook. CRC Press, Boca Raton, p 23/1
96. Matinlinna JP, Dahl JE, Lassila LVJ, Vallitu PK (2007) In: Mittal KL (ed) Silanes and other
coupling agents, vol 4. VSP/Brill, Leiden, p 82